Detection of faulty sensors in aero jet engine systems using robust model-based methods

Reviews the published state-of-the-art of research into fault diagnosis methods for jet engine systems using model-based (in contrast to knowledge-based) methods, with a particular emphasis on robustness with respect to modelling errors and unmeasurable disturbances. Reference is made to an important forum of activity in the field-an international technical cooperation panel which is investigating the evaluation of analysis algorithms for engine diagnosis. A brief description of a typical engine configuration and sensor set is then given. Results based on the application of one approach to model-based detection and isolation are presented. The detection algorithm is applied to a very complex thermodynamic system and the results illustrate very well the potential that a model-based method can offer when robustness to modelling errors (uncertainty) are correctly accounted for at the design stage